JP2018077457A - Method of performing autofocus, autofocus system, and camera comprising autofocus module - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress autofocus hunting when performing autofocus.SOLUTION: A method of autofocusing a camera during a zooming operation is provided, comprising (S1) zooming the lens to a first zoom position, (S2) measuring a first object distance from the camera to an object on which to focus, (S3) determining a first focus start position using the first object distance, (S4) performing a first autofocusing operation using the first focus start position as a starting point, thereby determining a first determined focus position, (S5) determining a first lookup object distance based on the first determined focus position, (S6) calculating first correction factor as a ratio between the first lookup object distance and the first object distance, (S13) zooming the lens to a second zoom position, and (S17) determining a second focus position using a second object distance (S16) calculated as a product of the first object distance and a second correction factor (S15) which is based on a ratio of depths of field at the second and first zoom positions.SELECTED DRAWING: Figure 6

Description

  The present invention relates to the field of camera autofocusing, and more particularly to autofocusing of PTZ cameras (ie, cameras capable of panning, tilting, and zooming).

  Autofocus is a function that is widely used in cameras. It is used in handheld cameras as well as monitor cameras, still image cameras and video cameras. Several different ways of performing autofocus are known and can generally be classified as either passive, active, or hybrid. One known passive method is called the maximum contrast method or contrast detection method. In this scheme, the contrast between adjacent pixels in the captured image is examined, and the focus position with the highest contrast is found and hopefully found by slightly changing the focus around the starting point. The search for maximum contrast can be performed, for example, using a mathematical optimization technique called “hill climbing”. Once the in-focus position is found, the in-focus position and zoom motor position are used to find a corresponding object distance trace curve in a set of trace curves provided by the lens designer or supplier. The trace curve plots the focus motor position against the zoom motor position, with one curve for each of several object distances. These curves are generally calculated theoretically using the theoretical characteristics of the various lenses included in the variable focus lens system. Several curves are provided, each based on a distinct object distance. When zooming in and out while keeping the focus at the same object distance, the trace curve is used to find the focus motor position that is supposed to give the proper focus. As a result, contrast determination need not be performed for each zoom position. If the object distance does not match any of the trace curves, interpolation is performed to obtain the object distance between the two curves.

  In some situations, such as in low lighting conditions or in a scene with low contrast or a point light source, it may be difficult to find the maximum contrast value, and the autofocusing algorithm will end in "hunting" In some cases, it means that the auto-focusing algorithm cannot find the maximum value and keeps changing the focus motor position.

  Even in good lighting conditions, when the algorithm is looking for the best focus motor position, focus changes that occur during the autofocus process may appear as annoying wobbling of the image.

  In the active autofocus method, an actual distance measurement to an object to be focused is performed using, for example, a laser. The in-focus position can then be determined using the lens trace curve. Such an active scheme requires the addition of a distance measurement system, which increases the cost of the camera but provides the advantage of facilitating autofocus in low illumination conditions. Such an active scheme also provides an immediate focus position advantage, so that there is no wobbling and “hunting” that may occur with a passive autofocus scheme.

  A challenge common to both passive and active methods is that the lens attached to the camera may be off the lens on which the trace curve was created, so that proper focusing is achieved for the determined object distance. Is not achieved using the in-focus position indicated. Individual lenses can be tested during camera manufacture to calculate the production offset from the theoretical trace curve to be used when selecting an appropriate focus position. However, such an offset test is usually performed only for the two zoom positions, namely “tele” and “wide” and only for focusing towards infinity. Furthermore, for reasons of production economy, in some cases, the offset test is only performed once per lens type and not for each individual lens. Even when each individual lens is tested, internal design tolerances and potential adjustments by the lens manufacturer will usually result in a practical trace curve shape that deviates from the theoretical curve. The production offset is not applied globally along the trace curve. The actual behavior of another lens specimen at other zoom positions and object distances may deviate significantly from the trace curve, even taking into account the factory or production offset. To make matters worse, the behavior of the lens may change over time, for example, due to temperature changes or due to changes in lens components over time.

  In order to improve autofocusing, a hybrid method combining a passive method and an active method can be used. For example, laser distance measurement can be used to provide a starting point for the contrast detection scheme, thereby facilitating autofocusing in highly challenging scenes. In addition, the hybrid method reduces the risk of “hunting” and performs the passive method more quickly with a potentially better starting point, thereby frustrating wobbling during the search for the optimal focus position Can be reduced.

  Problems associated with known autofocusing methods are often exacerbated by PTZ cameras. When the camera is panned and / or tilted in a new direction, autofocusing is necessary to focus on objects in the new field of view. In general, the autofocusing process is not fast enough to provide a focused image while panning and / or tilting from one point of the scene being monitored to another. Therefore, a focused image can be obtained at the start of movement and at the end of movement, but the image between them becomes blurred. The same is true for situations where the camera is zoomed in or out. This blurring during movement and zooming, for example, from a so-called “guard tour”, ie prior to panning, tilting and zooming operations performed to monitor various parts of the monitoring area. It may irritate an operator who is viewing a captured video sequence from a programmed sequence. Important visual information may be lost to the operator during panning, tilting, and zooming.

  An object of the present invention is to provide an auto-focusing method capable of supplying a better focused image. It is a further object of the present invention to provide an autofocusing method that can provide a better focused image during panning, tilting and zooming. Another object of the present invention is to provide a method of autofocusing that can be performed more quickly than prior art methods.

  According to a first aspect, these and other objects are achieved in whole or at least in part by a method of autofocusing a camera during a zooming operation, the camera comprising a zoom lens, the method comprising: Zooming to a first zoom position, measuring a first object distance from the camera to the object to be focused, and using the first object distance to determine a first focus start position Using the passive autofocus algorithm and using the first focus start position as a starting point of the algorithm to perform a first autofocusing operation, thereby Determining a focused focus position and determining a first lookup object distance based on the first determined focus position; Calculating a first correction factor as a ratio between the first lookup object distance and the first object distance, and determining a first depth of field at the first zoom position. Zooming the lens to a second zoom position; determining a second depth of field at the second zoom position; a first correction factor; and a second depth of field. Calculating the second correction factor as the product of the ratio between the first object depth and the first depth of field, and the second object distance as the product of the first object distance and the second correction factor. Calculating and determining a second focus position using the second object distance.

  Using such an autofocus scheme, it is possible to better utilize the trace curve provided by the lens manufacturer. Even if the individual lens used in the camera does not behave exactly as the lens for which the trace curve has been calculated, and the production offset determined for the lens type during camera manufacture is on the individual lens More suitable interpolation during zooming by calculating a first correction factor based on the measured object distance and the distance corresponding to the in-focus motor position found by the passive autofocus algorithm even if it does not necessarily fit It is possible to follow the measured object distance trace curve. In this way, a better focused image can be provided during zooming. Thus, an operator viewing a video sequence from the camera can obtain more information from the image and can be freed from the irritation of viewing the blurred image. A better starting point can speed up the focus search using the passive autofocus algorithm, so the operator is further freed from viewing the wobbling image during autofocusing using the passive autofocus algorithm. obtain.

  In addition, based on the ratio of the depth of field at the current zoom position to the depth of field at the previous zoom position during zooming, the update correction factor continues to be calculated, for example, once per captured or transmitted image frame. This can better account for deviations from the theoretical trace curve, so that for each frame or new zoom position, a new object distance to be used in the trace curve is determined.

  Once the lens has reached the desired zoom position, the second determined focus position using the focus motor position found using the trace curve during zooming as the starting point of the passive autofocus algorithm. Can be determined. Thereby, it is possible to find an appropriate focus more quickly. In addition, this is beneficial even when a guard tour is in progress but the operator interrupts the zooming operation by looking at something of interest in the scene. The latest focus motor position found using the correction factor and the trace curve can then be used as a starting point for the passive autofocus algorithm, resulting in a quick in-focus image. .

  A second lookup object distance can be determined based on the second determined in-focus position, and a new correction factor is calculated as a ratio between the second lookup object distance and the first object distance. be able to. In this way, an update correction factor that makes it easier to find an appropriate focus can be determined.

  The method includes zooming the lens to an intermediate zoom position between a first zoom position and a second zoom position, determining an intermediate depth of field at the intermediate zoom position, and a first correction. Calculating the intermediate correction factor as the product of the coefficient and the ratio between the intermediate depth of field and the first depth of field and the intermediate object as the product of the first object distance and the intermediate correction factor The method may further include calculating the distance and determining the intermediate focus position using the intermediate object distance. This allows the autofocus position to be determined at the intermediate zoom position without having to perform a time consuming passive autofocus algorithm. With such an approach, the correction factor can be calculated continuously during the zooming operation, so that good focus can be obtained by following a suitable trace curve.

  According to a variant, the method performs an intermediate autofocusing operation using a passive autofocus algorithm and using the intermediate focus position as the starting point of the algorithm, thereby providing an intermediate determined focus position. Further including determining. In this way, improved focus can also be achieved at the zoom position during the zoom operation from the first zoom position to the second zoom position. If you have a high input frame rate to the passive autofocus algorithm and allow time, such as when the user requests a low output frame rate, use the passive autofocus algorithm with the latest determined focus position as the starting point. , Thereby allowing possibly better focus than simply following the trace curve. This can be done, for example, at regular or irregular intervals rather than every frame during the zooming operation.

  Each time a passive autofocus algorithm is used, the correction factor can be updated. For example, the intermediate lookup object distance can be determined based on the intermediate determined focus position, and the new intermediate correction factor can be calculated as a ratio between the intermediate lookup object distance and the first object distance. The passive autofocus algorithm can work for reasons other than reaching a new zoom position. For example, a passive autofocus algorithm can operate when the camera pans or tilts, when the monitored scene changes, for example, due to lighting changes, and when the ambient temperature changes. If the object to be focused is at another distance other than the distance for which the latest stored correction factor is determined, the latest stored correction factor can still be used.

  The depth of field for each zoom position can be determined by a request from a lens controller that controls the lens.

  For many lenses, it may be beneficial to use different correction factors on either side of the transient zoom position. Therefore, the method may further include using a previously stored correction factor at a predetermined transient zoom position, wherein the previously stored correction factor is the first zoom position. Instead of being based on the depth of field of the zoom position between the transient zoom position and the zoom position between the transient zoom position and a third zoom position beyond the second zoom position. Based on depth of field. For example, at a zoom position of 5 × zoom, there may be a “kink” on the curve indicating the relationship between the depth of field ratio and the object distance. To handle this, a correction factor previously determined for the “other side” zoom position of the transient zoom position is used. Thus, what can be referred to as the tele correction factor can be continuously updated during zooming on the “tele” side of the transient zoom position, and what can be referred to as the wide correction factor is what is referred to as the transient zoom factor. It can be continuously updated during zooming on the “wide” side of the position. Once zooming in one direction of either zooming in or zooming out reaches the transient zoom position, the latest determined correction factor for the zoom position on the other side of the transient zoom position is the object distance calculation. Should be used for.

  Measuring the object distance from the camera to the object to be focused can be performed using a distance measurement system provided in the camera. The distance measurement system can be, for example, a laser distance measurement system.

  According to a second aspect of the present invention, the above object is achieved in whole or at least in part by an autofocus system for autofocusing a camera during a zooming operation, the autofocus system comprising: A distance measuring system arranged to measure a first object distance from the object to be focused to and an autofocus module arranged to carry out the method of the first aspect. The autofocus system of the second aspect can be implemented in the same way as the method of the first aspect with advantages.

  According to a third aspect, these and other objects are achieved in whole or at least in part by a camera comprising a zoom lens and an autofocus module arranged to perform the method of the first aspect. The

  The camera can comprise an integrated object distance measurement system arranged to measure a first object distance from the camera to the object to be focused. The object distance measurement system can be, for example, a laser distance measurement system.

  The camera of the third aspect can generally be implemented similarly to the method of the first aspect with advantages.

  According to a fourth aspect, these and other objects are fully met by a computer program product comprising a computer-readable storage medium having instructions configured to perform the method of the first aspect when executed by a processor. Or at least partially achieved. The processor may be any type of processor, for example, a central control unit (CPU), graphics processing unit (GPU), custom-made processing device implemented in an integrated circuit, ASIC, FPGA, or logic circuit including discrete components. Can do.

  Further scope of the applicability of the present invention will become apparent from the detailed description given below. However, although the detailed description and specific examples illustrate preferred embodiments of the present invention, various changes and modifications within the scope of the invention should be apparent to those skilled in the art from this detailed description. So it should be understood that it is given merely as an illustration.

  Thus, it is to be understood that the invention is not limited to the specific components of the described device or the method steps described, as such devices and methods may vary. Further, it is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. As used herein and in the appended claims, the articles “a”, “an”, “the”, and “said” refer to contexts It should be noted that unless specifically indicated otherwise, it is intended to mean that there is one or more elements. Thus, for example, reference to “an object” or “the object” can include several objects and the like. Further, the word “comprising” does not exclude other elements or steps.

  The present invention will now be described in more detail, by way of example, with reference to the accompanying schematic drawings.

FIG. 3 shows a scene being monitored with a PTZ camera, where the object is at various distances from the camera. It is a figure which shows the sight of FIG. 1 shown from the top. It is a block diagram of the camera of FIG. It is a block diagram which shows the autofocus module of the camera shown in FIG. FIG. 2 illustrates an example of a set of trace curves for a camera such as the camera of FIG. It is a flowchart which shows the method of autofocusing.

  As can be seen in FIG. 1, the scene 1 is monitored using a camera 2. The scene 1 is shown from above in FIG.

  The sight 1 has a building 3 with three doors 4, 5, 6. Each door has a respective passage 7, 8, 9. There are three shrubs 10, 11, 12 outside the building 3.

  In this example, the camera 2 is a PTZ camera, that is, a camera capable of panning, tilting, and zooming. The same kind of function can be achieved by using a camera with a zoom lens and mounting it on a so-called PT head, ie a device that pans and tilts another fixed camera.

  The camera 2 captures an image of the scene 1 and encodes it and sends it as a video sequence, for example to a control center, where the video sequence is decoded, so that the operator views the video sequence live be able to. The video sequence may further be stored for later review. The camera 2 can further have an on-board storage, for example an SD card, and can store a video sequence for later retrieval.

  The field of view of the camera 2 can only monitor a part of the scene 1 shown in FIG. 1 at a time. By panning, tilting, and zooming the camera 2, the operator can monitor various parts of the scene 1. This can be done manually, for example using a joystick. It may also be done using a pre-programmed guard tour that visits several points of interest in the scene in a predetermined sequence and stays at each point of interest for a predetermined time. Each time the camera is panned and tilted to a new position, it is very likely that the object to be focused is at a different distance from the camera, so the camera may need to be refocused. For example, to monitor people entering or leaving the building 3 through the left door 4, the camera may first be directed to the left door 4 and focused on that door. The camera 2 can then be panned to the right to cover the central door 5 that is slightly further away from the camera 2 than the left door 4. Thus, the camera 2 will need to be refocused to obtain a properly focused image of people passing through the central door 5. Subsequently, the camera 2 can be panned further to the right to capture an image of the right door 6 that is further away from the camera 2. Again, the camera 2 will need to be refocused to provide a properly focused image of people passing through the right door 6. For each such refocusing, an autofocusing procedure is used. Similarly, every time the camera 2 zooms in or out, it needs to be refocused unless the focus motor is locked to the zoom motor, such as a confocal zoom lens, sometimes called a true zoom lens. There is. With variable focus lenses, the focus is lost during zooming and the lens must be refocused.

  Reference is now made to FIG. 3, which is a simplified block diagram showing some components of the camera 2. The camera 2 has a zoom lens 20 or, strictly speaking, a variable focus lens. Lens 20 includes a zoom motor and a focusing motor, none of which are clearly shown in the drawings, but both are well-known mechanisms of zoom lenses. Further, the camera 2 includes an image sensor 21 for capturing an image of a scene to be monitored, and a processing circuit 22 for processing the image and encoding it. The camera has a memory 23 for use in the processing circuit 22, a local storage 24 for storing images in the camera, such as an SD card, and a network interface 25 for transmitting images from the camera 2 to, for example, a control center. It has further. The camera 2 has a laser distance measuring system 26 for measuring the object distance from the camera to the object to be focused. In addition, the camera 2 has an autofocus module 30 arranged to perform autofocusing.

  The autofocus module 30 is schematically shown in FIG. It comprises a distance input module 31, a passive autofocus algorithm module 32, and a correction coefficient update module 33.

  With reference to FIGS. 1-6, the function of the camera 2, more specifically the function of the autofocus module 30, will now be described with the help of a usage scenario. It will be appreciated that the camera 2 forms an integrated system for autofocusing.

The operator is monitoring the scene 1. Initially, the camera 2 is directed to the left door 4 where the camera 2 is to be focused. The lens 20 is set to the first zoom position (step S1 in FIG. 6). The laser distance measurement system 26 measures the distance D1 from the camera 2 to the left door 4 (S2). In order to select a suitable focus motor position for the measured distance D1, the trace curve provided by the lens manufacturer can be examined. An example of a set of trace curves for a lens is shown in FIG. If the lens 20 behaves exactly according to the trace curve provided by the lens manufacturer, find the curve corresponding to the object distance D1 and find the vertical focus motor position corresponding to the current zoom position of the horizontal axis. Thus, it is possible to easily determine an appropriate focusing motor position. However, this will in most cases not give the optimum focus position, even if the production offset is taken into account, since each individual lens may deviate from the behavior of the trace curve. Therefore, a passive autofocus algorithm is further used, thereby making the autofocus scheme a hybrid scheme. In this example, the passive autofocus algorithm is a contrast detection algorithm. The autofocus position found in the trace curve based on the measured distance D1 is labeled as a first autofocus start position AF1 _start (S3), and a newly adjusted auto called the first determined autofocus position AF1. A focus position is determined using a contrast detection algorithm (S4). When executing the contrast detection algorithm, the first autofocus start position AF1 _start is used as the start point. The search for the motor focus position that provides the maximum contrast is a range around the first focus start position corresponding to the depth of field DOF at the current zoom position, centered on the first autofocus start position AF1 _start. That is, for example, it can be executed in a range of 80% of the depth of field. The depth of field DOF can be obtained from a lens controller (not shown) that controls the lens (S7). The focus motor position AF1 found by the contrast detection algorithm is often different from AF1 _start . By examining the trace curve, the first lookup object distance D1 ′ can be found (S5), which corresponds to the first determined focus position AF1. The first correction factor ACF1 is calculated as a ratio between the first lookup object distance D1 ′ and the first object distance D1 according to the following general formula (S6).

  Where ACF is the correction factor, D is the measured object distance, and D 'is the object distance found using the passive autofocus algorithm. When zooming in to or from the first zoom position, the first correction factor is used to find the correct trace curve at the new zoom position. However, even better correction can be achieved if the depth of field at each zoom position is taken into account.

Returning to the example where the operator monitors the left door 4, a person appears at the doorway and the operator wants to zoom in on the person's face and attempt to identify the person. Therefore, the operator may wish to zoom in from the first zoom position to the second position. Zooming will occur over several image frames, not immediately. Here, it is assumed that the intermediate zoom position is reached after the first zoom position, and that the additional zoom position is reached afterwards, and then the second zoom position is reached. One skilled in the art will appreciate that any number of additional zoom positions can be traversed during zooming from the first zoom position to the second zoom position. When the lens is zoomed to the intermediate zoom position (S8), the depth of field at the intermediate zoom position is determined (S9) and labeled as intermediate depth of field DOFi. The ratio between the intermediate depth of field and the depth of field at the first zoom position labeled first depth of field DOF1 is calculated and the correction factor is the first correction factor ACF1. Is updated by calculating the intermediate correction coefficient ACFi as the product of the depth of field ratio DOFi / DOF1 (S10). Therefore, the updated correction coefficient is calculated according to the following general formula.

  Here, the subscript k means the previous zoom position, and the subscript k + 1 means the current zoom position. Each time a new correction factor is calculated, the previous correction factor is overwritten.

The intermediate object distance Di is calculated as the product of the measured first object distance D1 and the intermediate correction coefficient ACFi (S11). The calculation of the object distance at the intermediate zoom position can be described using the following general formula:

  Here, the subscript k means the previous zoom position, the subscript k + 1 means the current zoom position, and the subscript 1 means the initial measurement.

  This intermediate object distance Di is used to find the intermediate focusing motor position AFi in the trace curve (S12). In this way, a focused image can be provided without having to perform time-consuming contrast detection even at intermediate zoom positions.

  The lens is then zoomed to the additional zoom position and the process at the intermediate zoom position is repeated. Accordingly, the depth of field DOFa at the additional zoom position is determined, and the ratio between this additional depth of field and the intermediate depth of field is calculated. The correction factor is updated by calculating the additional correction factor ACFa as the product of the intermediate correction factor ACFi and the depth of field ratio DOFa / DOFi.

  The additional object distance Da is calculated as the product of the measured first object distance D1 and the additional correction factor ACFi. This additional object distance Da is used to find the additional in-focus motor position AFa in the trace curve. In this way, good focus can be achieved even at the additional zoom position.

  When the second zoom position is reached (S13), the depth of field DOF2 of the second zoom position is determined (S14), the correction coefficient is the additional correction coefficient, and the second depth of field DOF2. Updated by calculating the product of the ratio with the additional depth of field DOFa (S15). The second object distance D2 is calculated as the product of the second correction coefficient ACF2 and the measured first object distance D1 (S16). This second object distance is used to find the second focus position AF2 in the trace curve (S17). Using this second focus position AF2 as a starting point, a contrast detection algorithm may be used once again to find a proper focus at the second zoom position. This will give a second determined autofocus position, which will correspond to a second lookup object distance D2 '. The correction factor can now be updated by calculating the new correction factor ACFnew as the ratio of the measured first object distance D1 and second lookup object distance D2 '. The new correction factor can then be stored and used later when zooming in and out.

  The zooming process has been described so far as going from the first zoom position to the second zoom position via the intermediate zoom position and the additional zoom position, but from the first zoom position to the second zoom position. It should be understood that the distance between the zoom positions is so short that there may be instances where there are no image frames between them with other zoom positions, such as an intermediate zoom position and an additional zoom position. In such a case, the method is such that the second object distance is the first object distance D1, the first correction factor ACF1, and the ratio DOF2 between the second depth of field and the first depth of field. Adapted to be calculated as a product of / DOF1. This can occur, for example, during very slow zooming at low output frame rates.

As stated in the Summary of the Invention, the lens may behave in a manner that requires the use of different correction factors in the “tele” range of the zoom position and the “wide” range of the zoom position. As an example, this transient zoom position may occur at 5x zoom. One correction factor may be used in a “tele” range called ACF _T, and another correction factor may be used in a “wide” range called ACF _W. Both of these correction factors are calculated in the same way as described above, but each is for zooming between zoom positions on each side of the transient zoom position.

Returning again to the example where the operator monitors the left door 4 of the scene 1, as just described, the camera has zoomed from the first zoom position to the second position, and the two zoom positions are Both are on the “tele” side of the transient zoom position. After examining the face of the person appearing at the doorway, the operator can zoom out to the third zoom position and monitor the left door 4 as well as the central door 5, but keep the distance to the object to be focused on. I want to do so. Zooming back from the second zoom position zoomed into the person's face to the first zoom position slightly wider but still on the “tele” side, the same process just described is the same during the zooming operation. Used to obtain an in-focus image. If you continue to zoom out, the lens 20 passes through the transient zoom position. In order to calculate the transient object distance Dt, the previously stored correction factor ACF _W for the “wide” side of the transient zoom position is used. The wide correction coefficient ACF _W is determined in the same manner as described above, but using the depth of field at the zoom position on the “wide” side of the transient zoom position.

The transient depth of field DOFt is determined and the wide correction factor ACF _W is updated by calculating the product of the previously stored wide correction factor and the depth of field ratio and is more general in the following formula: Can be described.

Similarly, when zooming in "tele" side of the transitional zoom position, tele correction factor ACF _T is updated as follows.

  As before, the subscript k means the previous zoom position, and the subscript k + 1 means the current zoom position.

Thus, transient object distance Dt is calculated as the product of the first object distance D1, which is measured and updated wide correction coefficient ACF _W. As before, the transient object distance Dt is used to find the transient focus motor position AFt in the trace curve.

When the camera 2 and the autofocus module 30 are started for the first time, the autofocus module 30 is initialized by setting the value of each correction coefficient to 1, that is, ACF _W = 1 and ACF _T = 1. As soon as a new correction factor is calculated by dividing the object distance determined with the contrast autofocus algorithm by the object distance measured with the laser distance measurement system, the corresponding correction factor is updated with the new value. Thus, if the lens is set to the “wide” range zoom position, the ACF _W is updated, while if the lens is set to the “tele” range zoom position, the ACF _T is updated. Similarly, each correction factor ACF _W or ACF _T is updated using the ratio between the depth of field at successive zoom positions as described above.

ここで、Ｎは開口設定値またはＦナンバであり、ｃは錯乱円であり、ｆは焦点距離であり、ｓは物体距離である。上述の説明では、焦点距離が詳細に論じられるのではなく、むしろズーム位置が論じられていることに留意することができる。しかしながら、これらの２つのパラメータは、ズームモータ位置の変化がレンズの焦点距離の変化を伴うので密接に関連している。 The lens controller can determine the depth of field according to the following formula at medium to large object distances:

Here, N is an aperture setting value or F number, c is a circle of confusion, f is a focal length, and s is an object distance. It can be noted that in the above description, the focal length is not discussed in detail, but rather the zoom position is discussed. However, these two parameters are closely related because changes in the zoom motor position are accompanied by changes in the focal length of the lens.

ここで、ｍは倍率である。 For close-up, the depth of field can be determined according to the following formula:

Here, m is a magnification.

  Different lens manufacturers may use slightly different formulas for depth of field calculations, but as long as the depth of field can be determined, the accuracy of the formula used is It should be noted that it is not important for the autofocusing scheme of the present invention.

  Those skilled in the art will appreciate that the embodiments described above can be modified in many ways and still benefit from the advantages of the present invention as shown in the embodiments described above. As an example, the method of the present invention has been described in the context of zooming to an object at a distance from the camera, but the same principle applies to objects that move away from or approach the camera as the object distance changes. Can be used for constant zoom. The same stored correction factor may be used regardless of the object distance, or one correction factor may be stored for each of several object distances. The choice of approach can depend on the type of lens used.

  Furthermore, in the above-described embodiments, the camera is equipped with a laser distance measuring system that is used to measure the distance from the camera to the object to be focused, but the measuring system can instead be e.g. Ultrasound can be used. It is also possible to use a separate measurement system that is not integrated in the camera. A tape measure or measuring rod can be used as a simple measuring system.

  Instead of actually measuring the distance to the object to be focused, another passive autofocus method can be used. For example, the same algorithm implemented with another camera, perhaps more reliable, can be used as a reference instead of a physically measured distance. Alternatively, another autofocus algorithm can be run on the same camera and used as a reference. This can be beneficial if other autofocus algorithms provide better results, but requires more time or computational power.

  As described above, the depth of field is determined when the camera reaches a new zoom position. However, the order in which the method steps are performed is not important. For example, in preparation for zooming from the first zoom position to the second zoom position, the depth of field at some intermediate zoom positions between the first zoom position and the second zoom position It may be requested by the lens controller as already known before the intermediate zoom position is reached.

  In the example described, the passive autofocus algorithm is a contrast determination algorithm. One skilled in the art will appreciate that other passive autofocus algorithms can be used. For example, a phase detection algorithm can be used.

  The correction factor ACF was first described as being calculated as the ratio between the object distance corresponding to the in-focus motor position obtained with the contrast detection algorithm and the object distance measured with the laser measurement system. Since the object distance and focus motor position are more closely linked by the trace curve, it is corrected as the ratio between the focus motor position obtained by the contrast detection algorithm and the focus motor position corresponding to the measured object distance. It should be understood that the coefficients can also be calculated.

  Trace curves from lens manufacturers can be implemented in an autofocusing manner in any desired way. For example, discrete points on the trace curve can be stored in a lookup table. At distances where the trace curve set does not have its own curve, the values can be interpolated. Alternatively, instead of forming a lookup table, a function in the form of a polynomial is fitted to each trace curve and used to determine the relationship between object distance, zoom motor position and focus motor position Also good.

  As an example of the transient zoom position, 5x zoom has been described. However, the transient zoom position can vary depending on the lens type. Depending on the lens, there may be more than one transient zoom position. For some lenses, it may be possible to continue to calculate updated correction factors without using previously stored correction factors for zoom positions beyond the transient zoom position. Depending on the lens type, there may even be no transient zoom position. The behavior of a particular lens type can be determined by testing.

  Although the present invention has been described in the context of a PTZ camera, as already mentioned, a fixed zoom camera can also function as a PTZ camera if it is placed on the PT head. Furthermore, the present invention is equally applicable to fixed zoom cameras, often referred to as fixed box cameras, that do not have panning and tilting capabilities, but have zoom lenses or, more strictly speaking, variable focus lenses.

  The camera can be a monitoring camera. It can be a digital camera or an analog camera. The camera can be any type of camera such as a visible light camera, an IR camera, or a thermal camera.

  As described above, the camera 2 forms an integrated system for autofocusing. However, such a system may also be composed of two or more separate components. For example, the distance measurement system need not be integrated into the camera and may be provided separately. The object distance measured by the measurement system is then supplied to the autofocus module for use in the autofocus method. Further, the autofocus module may be disposed outside the camera and operably connected to the camera. As a further option, the distance measurement system and the autofocus module may be located in a common unit and operably connected to the camera.

  Using the concepts of the present invention, faster autofocusing can be performed during zooming operations or when the object to be focused is away from or approaching the camera at a constant zoom position. Is possible. In this way, when the passive autofocus algorithm is executed, there is less wobbling and better focus can be achieved for frames during use of the passive autofocus algorithm. Since the correction factor is continuously updated, it is possible to compensate for changes in lens behavior caused by, for example, ambient temperature changes or lens aging. Further, the correction factor can be monitored and an alarm event can be generated if one or both of the correction factors is greater than a predetermined threshold. This can indicate that the lens needs to be recalibrated, serviced, or further replaced.

  Accordingly, the invention should not be limited to the illustrated embodiments, but should be defined only in accordance with the appended claims.

Claims (12)

A method for autofocusing a camera during a zooming operation, the camera comprising a zoom lens, the method comprising:
Zooming the lens to a first zoom position;
Measuring a first object distance from the camera to the object to be focused;
Determining a first focus start position using the first object distance and using a relationship between the object distance, the zoom position and the focus position;
A first autofocusing operation is performed using a passive autofocus algorithm and using the first focus start position as the start point of the algorithm, thereby determining a first determined focus position. To decide,
Determining a first lookup object distance based on the first determined focus position;
Calculating a first correction factor as a ratio between the first lookup object distance and the first object distance;
Determining a first depth of field at the first zoom position;
Zooming the lens to a second zoom position;
Determining a second depth of field at the second zoom position;
Calculating a second correction factor as a product of the first correction factor and a ratio between the second depth of field and the first depth of field;
Calculating a second object distance as a product of the first object distance and the second correction factor;
Using the second object distance and determining the second focus position using the relationship between the object distance, the zoom position and the focus position. A second autofocusing operation is performed using the passive autofocus algorithm and the second focus position as a starting point of the algorithm, thereby determining a second determined focus position. To decide,
Determining a second lookup object distance based on the second determined focus position;
The method of claim 1, further comprising calculating a new correction factor as a ratio between the second lookup object distance and the first object distance. Zooming the lens to an intermediate zoom position between the first zoom position and the second zoom position;
Determining an intermediate depth of field at the intermediate zoom position;
Calculating an intermediate correction factor as a product of the first correction factor and a ratio between the intermediate depth of field and the first depth of field;
Calculating an intermediate object distance as a product of the first object distance and the intermediate correction factor;
The method of claim 1, further comprising determining an intermediate focus position using the intermediate object distance. Further comprising performing an intermediate autofocusing operation using the passive autofocus algorithm and using the intermediate focus position as a starting point of the algorithm, thereby determining an intermediate determined focus position. The method according to claim 3. Determining an intermediate lookup object distance based on the intermediate determined focus position, and calculating a new intermediate correction factor as a ratio between the intermediate lookup object distance and the first object distance The method of claim 4 comprising.   The method according to claim 1, wherein a depth of field is determined by a request from a lens controller that controls the lens. At a given transient zoom position, the previously stored correction factor is used,
Instead of the previously stored correction factor being based on the depth of field of the zoom position between the first zoom position and the transient zoom position, the transient zoom position and the second zoom 7. A method according to any one of the preceding claims, wherein the method is based on a depth of field at a zoom position between a third zoom position beyond the position.   The measurement of the first object distance from the camera to the object to be focused is performed using a distance measurement system provided in the camera. The method according to item. An autofocus system for autofocusing a camera during zooming operation,
A camera,
A distance measuring system arranged to measure a first object distance from the camera to the object to be focused;
An autofocus system comprising an autofocus module arranged to perform the method according to any one of claims 1-8.   A camera comprising a zoom lens and an autofocus module arranged to carry out the method according to claim 1. The camera of claim 10, further comprising an integrated object distance measurement system arranged to measure the first object distance from the camera to the object to be focused.   A computer program product comprising a computer readable storage medium having instructions configured to perform the method of any one of claims 1 to 8 when executed by a processor.

Images